Autoimmune pancytopenia

Autoimmune Pancytopenia Lymphocyte Inhibition of Autologous But Not Allogenic Bone Marrow Growth in Vitro

G. DAVID ROODMAN, M.D., Ph.D.* JOAO L. ASCENSAO, M.D. MUSSA BANISADRE, M.D. PHILLIP M. BLOOM, M.D. ESMAIL D. ZANJANI, Ph.D. Minneapolis, Minnesota

From the Veterans Administration Medical Center, Regional Kidney Disease Program at Hennepin County Medical Center and the Departments of Medicine and Physiology, University of Minnesota School of Medicine, Minneapolis, Minnesota. This study was supported by the Veterans Administration and by NIH Grants CA-23021 and AM-24027. Requests for reprints should be addressed to Dr. Esmail D. Zanjani, Veterans Administration Medical Center (151). 84th Street & 48th Avenue South, Minneapolis, Minnesota 55417. Manuscript accepted January 10,198O. * Present address: University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Drive, San Antonio, Texas 78284.

A patient with autoimmune renal failure, cavitary lung lesions and arthritis experienced pancytopenia while prednisone therapy was being tapered. Utilizing semisolid culture techniques, a population of nonadherent peripheral blood mononuclear cells was demonstrated, which inhibited autologous but not allogeneic bone marrow erythroid colony-forming units (CFU-E)and myeloid colony-forming units (CFU-c) in vitro. No inhibition of CPU-E or CFU-c colony formation was seen when patient’s serum or immunoglobulin G (IgG) was added to cultures. Reinstitution of prednisone therapy resulted in normalization of peripheral blood counts, which was accompanied by the loss of the hemopoietic inhibitor cell activity in the patient’s peripheral blood. These results demonstrate the need for testing autologous marrow samples when looking for possible immunemediated inhibition of hematopoiesis. Immune suppression of hematopoiesis has been suggested as a possible mechanism for the pancytopenia in some patients with aplastic anemia, and other hematologic and rheumatic disorders [l]. Lymphocyte suppression of erythropoiesis [2-51 and myelopoiesis [6-81 has been demonstrated in a variety of patients. Antibody-mediated suppression of hematopoiesis has also been implicated in these disorders [9,10]. Generally, these findings have been based on in vitro co-culture studies involving allogeneic effector-target cell combinations [3,6]. It has been argued that such a system may not represent an appropriate approach since it is possible that co-culture studies employing allogeneic target cells may fail to reveal the presence of an immune-mediated mechanism in a subpopulation of these patients. We report studies carried out in a patient with autoimmune renal failure, cavitary lung lesions and arthritis who experienced pancytopenia while prednisone therapy was being tapered. Utilizing semisolid culture techniques, we have shown that lymphocytes in the patient’s peripheral blood mononuclear cells inhibited erythroid and granulocyte-macrophage colony formation in the patient’s, but not in normal, human bone marrow cells. With reinstitution of prednisone therapy the patient’s hematologic parameters normalized, and the hematopoiesis-inhibitory activity of the patient’s peripheral blood lymphocytes was no longer detectable. CASE REPORT A 27 year old man presented to another hospital in December 1973 with migratory polyarthritis, hematuria, proteinuria, eosinophilia and a palpable

August 1990 The American Journal of Medicine Volume 69

325

AUTOIMMUNITY,

PANCYTOPENIA

AND AUTOLOGOUS MARROW SUPPRESSION-ROODMAN

spleen. Renal biopsy revealed a focal necrotizing glomerulitis. A diagnosis of vasculitis was considered and therapy with prednisone (60 mg daily) was started and subsequently reduced to 40 mg every other day. In June 1974 the patient presented with recurring polyarthritis, fever, hematuria. proteinuria, hemoptysis, dyspnea, a confluent purpuric macular rash and tender nodules on his fingers and elbows. The findings on a repeat renal biopsy were essentially unchanged. A transbronchial lung biopsy specimen showed hemosiderosis. The dose of prednisone was increased to 40 mg daily without improvement. In September 1974 the patient was noted to have pulmonary infiltrates with cavitation which resolved spontaneously. A clinical diagnosis of Wegener’s granulomatosis was made in June 1975, and therapy with cyclophosphamide (200 mg daily) was started and prednisone was continued. In December prednisone therapy was discontinued, and the cyclophosphamide was decreased to 150 mg daily. In November 1977 the patient experienced hemorrhagic cystitis and cyclophosphamide therapy was stopped. In July 1978 the patient was noted to have leukopenia with white blood cell counts of 1,500 to 2,00O/mms, which persisted through 1978. In January 1979 the patient was hospitalized and was again noted to have splenomegaly. Laboratory studies revealed a hemoglobin level of 12.6 g/dl, a white blood cell count of 1,500/mm3 with 37 percent neutrophils, a decreased haptoglobin level and a platelet count of 83,000/mm3. A bone marrow biopsy specimen was hypercellular with normal distribution of marrow elements. Epinephrine stimulation revealed no significant white blood cell margination, but a single test dose of prednisone resulted in a significant increase in the white blood cell count. On February 1.1979. the patient was readmitted. On admission, the serum creatinine level was 1.9 mg/lOO ml and the creatinine clearance 70 cc/min. Within 18 days his creatinine level rose to 20 mg/lOO ml and his creatinine clearance fell to 7 cc/min, for which therapy with prednisone (60 mg daily] and cyclophosphamide (150 mg daily) was reinstituted. The patient was referred to the Hennepin County Medical Center for further evaluation and initiation of hemodialysis. During immunosuppressive therapy, the white blood cell count ranged from 4,000 to 5,OOO/mms with a normal differential. To better define the activity of the patient’s disease, cyclophosphamide therapy was discontinued and the dose of prednisone was tapered over 30 days. By the end of May the patient was receiving 5 mg alternating with 2.5 mg/day of prednisone. The patient’s peripheral blood cell counts steadily decreased until, on June 5th, the white blood cell count was l,OOO/mms, platelet count 100,000/mm3. hemogloblin level 5 g/dl, reticulocyte count 1.3 percent, the plasma hemoglobin level normal and the haptoglobin level low normal. No evidence of splenic sequestration was found. A bone marrow aspirate and biopsy, performed on June 5th. revealed only hypocellular marrow. In vitro studies of the patient’s bone marrow and peripheral blood were performed at that time. Subsequently, prednisone therapy was reinstituted (20 mg/day) and the patient was given a transfusion. By June 19.1979, the patient’s hemoglobin level was stable at 10 g/dl. the white blood cell count was 8,000/ mms, with a normal differential and the platelet count was 21O,OOO/mms. In vitro assays of peripheral blood and bone marrow were repeated at this time. The patient’s peripheral blood counts have remained stable. He is currently undergoing

326

August 1980

The American Journal of Medicine

hemodialysis tation. MATERIALS

and is being AND

ET AL.

evaluated

for renal

transplan-

METHODS

Peripheral Blood and Bone Marrow Samples. Peripheral blood and bone marrow aspirates were obtained from the patient and concommitantly from controls when the patient was pancytopenic and after reinstitution of prednisone therapy. Informed consent was obtained prior to each procedure. Peripheral blood and marrow samples were processed as previously described [ll]. Briefly, peripheral blood and marrow were collected in heparinized syringes, diluted 1 to 1 with alpha minimal essential medium ((UMEM) and layered over Hypaque-ficoll gradient. The mononuclear cells were collected from the interface, washed three times, resuspended in aMEM and counted. Peripheral blood mononuclear cells were either used directly in co-culture experiments or were incubated in oMEM supplemented with 20 percent fetal calf serum at 37% for 1 hour in plastic culture dishes. The nonadherent cells were collected, washed three times in aMEM and used in co-culture studies. Since the patient was undergoing hemodialysis and had received a transfusion, peripheral blood mononuclear cells were also prepared from blood of a patient undergoing hemodialysis whose blood transfusion requirements had been similar to those of our patient. Serum Samples and Immunoglobulin G (IgG). Twenty milliliters of serum before and after dialysis were obtained from the patient, another patient undergoing dialysis who had received similar transfusions and the normal control. Serums were added to bone marrow culture at three different concentrations (0.05,O.l and 0.2 ml/l.1 ml of total culture volume). IgG was prepared from each of the serum samples according to the method of Reif [12] and reconstituted in phosphatebuffered saline solution at the same concentration as present in initial serum samples. Bone Marrow Cultures. Erythroid colony-forming units (CFU-E): Bone marrow cells were cultured for seven days at 37% in an atmosphere of 5 percent carbon dioxide (CO&air in the plasma clot with or without erythropoietin 1.5 IU/l.l ml of culture according to the method of Tepperman et al. [13]; 2 to 6 X lo5 marrow cells were used in all cultures. For studies employing peripheral blood mononuclear cells, 6 x lo5 marrow cells were co-cultured with either 2 or 4 x lo5 peripheral blood mononuclear cells or peripheral blood lymphocytes. All results were tabulated per 6 X lo4 marrow cells cultured and represent the mean f standard error of the mean of quadruplicate samples. Only colonies containing 8 or more benzidine positive cells were scored. Granulocyte-macrophage colony-forming units (CFUc): The assay for CFLJ-c was performed as described before [6]. Human peripheral blood leukocytes obtained by dextran sedimentation were used as feeder layers at a concentration of 1 x 106 cells in 0.5 percent agar/plate. Two hundred thousand cells or equal mixtures of patient and control cells were cultured in quadruplicate in 0.3 percent agar in McCoy’s medium supplemented with fetal calf serum for seven days, after an overnight incubation, at 37% in 7.5 percent COa-air. Cultures were scanned for the presence of colonies (>40 cells) and clusters (3 to 40 cells) under a dissecting microscope. Statistical Methods. Student’s t test was used to evaluate

Volume 69

AIJTOIhlMUNI’I’Y.

TABLE I

PANCYTOPEKIA

AND A1ITOl.O(;OUS

Effects of Peripheral Blood Lymphocytes on CFU-E and CFU-c Derlved Colony Formatlon Prior to Prednlsone Therapy

Additions to Cutture None Normal peripheral blood lymphocytes Dialysis control peripheral blood lymphocytes Patient peripheral blood lymphocytes

Nonal Marrow CFU-E CFU-c 295 f 259f

12 79f 5 10 71 f3

228f20

70f5

339 f

16 91 f 5

Patient Marrow CFU-c CFU-E 77 f 90f6

10

92flO

28f

22 f 7 19 f 2

19f2

4

5&l

NOTE: Results are presented as CFU-E derived erythroid colonies or total CFU-c derived aggregates (colonies and clusters) per 6 X 10’ marrow cells cultured. Each value represents the mean f SEM for quadruplicate determinations.

differences

Differences

in results from quadruplicate determinations. with p
RESULTS

From the data presented here one can postulate a relntionship between the cessation of prednisone therapy and the onset of pancytopenia. Reinstitution of the prednisone therapy led to a significant improvement in the peripheral blood counts. Results from in vitro coculture studies of marrow samples taken at the intervals indicated are presented in Tables I, II, III and IV. The patient’s bone marrow, obtained during the cytopenic phase. produced significantly fewer erythroid and granulocyte-macrophage colonies than normal human bone marrow cultured simultaneously (p
TABLE Ill

Effects of PBL on CFU-E and CFU-c Derlved Colony Formatlon After Two Weeks of Prednlsone Therapy

Addlttonsto Culture No PBL Control PBL Patient PBL

NormalMarrow CFU-E CFU-c 288 f 26 249 f 9 339f25

106 f 30 96f 34 156f64

MARROW

181 f 10 30 f 1 160f 15 33 f 3 155f25 31f3

NOTE: PBL = peripheral blood lymphocytes. Results are presented as colonies per 6 X 10’ marrow cells cultured. Each value represents the mean f SEM for quadruplicate determinations. Cultures for CFU-E contained 1.5 U/ml erythropoietin, and for CFU-c contained a source of colony stimulating factor.

ET AL.

Effect of PBL and PBMC on CFU-E and CFU-c Derlved Colony Formation Prior to Prednlsone Therapy

TABLE II

No. ot Colonles CFU-c CFU-E

Addttionsto Culture None PBMC PBL

77 f 10 16 f 2 28 f 4

22 f 7 4fl 5fl

NOTE: PBMC = peripheral blood mononuclear cells, PBL = peripheral blood lymphocytes. Results are presented as number of colonies formed per 6 X 10’ marrow cells cultured. Each value represents the mean f SEM of quadruplicate determination. All cultures contained 1.5 U/ml erythropoietin for CFU-E or a source of colony stimulating factor for CFU-c.

received similar transfusions, and a normal volunteer failed to significantly inhibit or enhance (p >0.05) the patient’s marrow or the normal marrow in vitro [Table I]. The presence of adherent cells in the patient’s peripheral blood mononuclear cells did not significantly change the results (p >0.05) (Table II). Two weeks after reinstitution of prednisone therapy, no inhibiting activity was detectable with co-culture of the patient’s peripheral blood lymphocytes and his bone marrow (Table III). Serum and IgG from the patient, a control patient undergoing dialysis and a normal volunteer did not inhihit the patient’s marrow CFU-E or the normal marrow CFU-E at the concentrations tested, when the patient was pancytopenic (Table IV]. COMMENTS

IJsing semisolid culture techniques we have shown cell-mediated suppression of autologous bone marrow crythroid (CFU-E) and CFU-c colony formation. Allogeneic co-culture experiments did not demonstrate hematopoietic inhibitory activity in this system. The inhibitor cell population was found in the nonadherent cell fraction of the patient’s PBMC, suggesting that it is TABLE IV

Effects of Serum and Human IgG on CFU-E Derlved Colony Formatlon Prior to Prednlsone Therapy

Additionsto Culture

Patient Marrow CFUE CFU-c

SIJPPRESSION-RO0nh~AN

None Dialysis serum Normal serum Patient’s serum Dialysis IgG Normal IgG Patient’s IgG

NormalMarrow 295 210 323 334 353 211 220

f f f f f f f

12 17 18 24 47 39 51

Patknt’s Marrow 77f 10 62 f 6 72 f 8 82f 10 65 f 12 93f 10 72 f 8

NOTE: Results are presented as erytkokt cotonii per 6 X 10’ bone marrow cells cultured. Each value represents the mean f SEM for all concentrations of serum or IgG tested in quadruplicate determinations. Dialysis serum and IgG refer to predialysis control serum or IgG of a patient undergoing hemodialysis with a similar degree of renal failure and transfusion history.

August 1980 The American Journal of Medicine

Volume 69

327

AUTOIMMUNITY. PANCYTOPENIA AND AUTOLOGOUS MARROW SUPPRESSION-ROODMAN

a lymphocyte. In patients with aplastic anemia, it has been suggested that transfusion-induced alloimmunization may induce inhibitor lymphocytes detectable in co-culture experiments [14], although it is currently recognized that patients with aplastic anemia given no transfusions, may also have inhibitor cell activity [15]. The inhibitor cell population from our patient failed to suppress colony formation in normal marrow while markedly inhibiting it in his own marrow. Peripheral blood lymphocytes from a patient undergoing dialysis, who received the same number of transfusions as our patient, had no inhibitory activity. Therefore, it seems unlikely that the inhibitor cell activity demonstrated in our patient was caused by alloimmunization. The effects of this inhibitor cell population are similar to those reported by Kagan et al. [16] who showed autologous but not allogenic inhibition of CFU-c in patients with systemic lupus erythematosus. We cannot state with certainty that our in vitro studies reflect the in vivo situation. However, evidence suggests that the inhibitor cell was involved in the pathogenesis of this patient’s pancytopenia: (1) the inhibitor cell affected autologous but not allogeneic CFU-E and CFU-c, and (2) results of the clinical and in vitro culture studies correlated with the reinstitution of glucocorticoid ther-

ET AL.

apy. It seems unlikely that this inhibitory activity was exercised at the level of the pluripotent stem cell since the patient’s peripheral blood cell counts normalized within two weeks after starting prednisone therapy. Possibly, this autologous inhibitor activity was directed against the patient’s HLA or Ia antigens which are shared by all hemopoietic cells [17]. The inhibitor cell in our patient is a nonadherent glucocorticoid-sensitive peripheral blood mononuclear cell. Several investigators have shown that the inhibitor cell in aplastic anemia [1.2,6] and immune neutropenia [7,18] may be a T lymphocyte. Furthermore, a number of cytotoxic [18] and suppressor functions of T cells [ZO]are inhibited by glucocorticoids. This suggests that the inhibitor cell in our patient may be a T lymphocyte. However, T and B cell fractionation of the patient’s peripheral blood was not undertaken because of the patient’s marked leukopenia. It is important to reiterate the need for testing autologous inhibition in such patients which might otherwise be missed by co-culture experiments using only normal marrow. ACKNOWLEDGMENT We wish to thank Ms. T. Engler and Ms. A. McHale for excellent technical assistance.

REFERENCES 1.

2.

3. 4. 5. 6. 7. 0.

9.

10.

11.

328

Cline Ml. Golde DW: Immune suooression in hematoooiesis. Am J,Med 1978: 64: 304. .. Hoffman R. Kopel S, Hsu SD, et al.: T cell chronic lymphocytic leukemia: presence in bone marrow and peripheral blood of cells that suppress ervthropoiesis in vitro. Blood 1978; 52: 255. ._ . Hoffman R, Zanjani ED, Lutton JD. et al.: Suppression of erythroid-colony formation by lymphocytes from patients with aplastic anemia. N Engl J Med 1977: 296: 10. Hoffman R. Zanjani ED, Vila J, et al.: Diamond-Blackfan syndrome: lymphocyte-mediated suppression of erythroooiesis. Science 1976: 193: 899. Liiwin SD, Zanjani ED: Lymphocytes suppressing both immunoglobulin production and erythroid differentiation in hypogammaglobulinaemia. Nature 1977; 266: 57. Ascensao IA. Pahwa RN, Kagan WA, et al.: Aplastic anemia: evidence for an immunologic mechanism.iancet 1976: 1: 669. Kagan WA, Ascensao JL, Fialk MS. et al.: Studies on the pathogenesis of aplastic anemia. Am 1 Med 1979: 66: 444. Bagby GC. Jr. Gabourel JD: Neutropenia in three patients with rheumatic disorders. Suppression of granulopoiesis by cortisol-sensitive thymus-dependent lymphocytes. J Clin Invest 1979: 64: 72. Hoffman R, Dainiak N. Sibrack L. et al.: Antibody-mediated aplastic anemia and diffuse fascitis. N Engl J Med 1979; 300: 718. Blaschke 1. Goeken NE. Thomoson IS. et al.: Acauired agranul&ytosis with granulocyte specific cytotoxic’autoantibody. Am J Med 1979; 66: 862. Rinehart JJ, Zanjani ED, Nomdedeu 1~.et al.: Cell-cell in-

August 1980

The American Journal of Medicine

12. 13. 14.

15. 16. 17. 18. 19. 20.

teraction in erythropoiesis. Role of human monocytes. J Clin Invest 62: 979,1979. Reif A: Batch preparation of rabbit y G-globulin with DEAE cellulose. Immunochemistry 1969; 6: 723. Tepperman AD, Curtis JE. McCulloch EA: Erythropoietic colonies in cultures of human marrow. Blood 1974; 44: 659. Singer JW. Brown JE, James MC, et al.: Effect of peripheral blood lvmohocvtes from oatients with aolastic anemia on granulocyiic colony growth from HLA-matched and mismatched marrows: effect of transfusion sensitization. Blood 1978: 52: 37. Singer JW. Doney KE, Thomas ED: Co-culture studies of 16 untransfused patients with aplastic anemia. Blood 1979: 54: 180. Kagan WA, Weintraub LR. Cathcart E. et al.: Neutropenia in systemic lupus erythematosus. Blood 1979; [suppll] 64: 155a. Schlossman SF. Chess L. Humohrevs RE. et al.: Distribution of la-like molecules on the surface of normal and leukemic human cells. Proc Nat] Acad Sci USA 1976; 73: 1288. Abdou N. Napombejara C, Balentine L. et al.: Suppressor cell-mediated neutropenia in Felty’s syndrome. 1Clin Invest 1978: 61: 738. Stavy L. Cohen IR. Feldman M: The effects of hydrocortisone on lymphocyte mediated cytolysis. Cell lmmunol 1973: 7: 302. Saxon A, Stevens RH, Ramer SJ. et al.: Glucocorticoids administered in vivo inhibit human suooressor T-lvmohocvte function and diminish B-lymphocyie responsiveness in in vitro immunodobdin svnthesis. 1 Clin Invest 1978; 61: 922.

Volume 69